Viral vector particles such as adenovirus gene transfer vectors are used in vitro and in vivo in order to transfer genes into living cells for therapeutic purposes (gene therapy), for vaccination purposes and for functional studies. Adenoviral vector particles can be produced in high titers, their genome is highly stable and they have the capability to transduce proliferating and resting cells. One can distinguish between different types of adenovirus vector particles: for example there are so-called first generation vector particles (the E1 functions are deleted), second generation vector particles (additional deletion of E2 and/or E4 functions), high-capacity vector particles (deletion of most or all viral genes). Furthermore, there are also replication-competent adenovirus vector particles (commonly referred to as oncolytic adenoviruses), which are for example used in tumor therapy. Additionally there are also chimeric adenovirus vector particles, whereby capsids of different serotypes are being combined. One example is the substitution of the natural serotype 5 Fiber protein with a Fiber protein derived from a different serotype such as for example serotype 17 or serotype 9.
Adenoviruses were found in numerous species. More than 50 different human serotypes are known, for example Ad12 (Subgenus A), Ad3 and Ad7 (Subgenus B), Ad2 and Ad5 (Subgenus C), Ad8 (Subgenus D), Ad4 (Subgenus E), Ad40 (Subgenus F) (Wigand, 1986). Apart from being detected in human beings, adenoviruses were also found in most vertebrates including chimpanzees, cattle, pigs, mice and chicken. Most of the currently used vector particles for gene transfer are based on adenovirus type 5 (Ad5), although numerous laboratories are also working on the development of vector particles which are either based on a different human-pathogenic serotype or which were isolated from chimpanzees, cattle or another species.
The transduction of cells by Ad5 vector particles takes place via at least two receptors. Initially the vector particles bind with the Knob domain of the capsid protein Fiber to the Coxsackie and Adenovirus receptor (CAR). This interaction induces minor structural modifications in the vector particle which allow an interaction of the penton base protein of the vector particles with the integrins of the cell surface. Finally, there is an integrin-mediated uptake of the vector particles into the cell via receptor-mediated endocytosis. During uptake of the vector particles and in the early endosome a regulated, gradual disassembly of the vector particles takes place by substantial participation of the viral cysteine protease p23 and which is completed after leaving the early endosome and upon entering the nucleus. This process depends on the redox status of the cysteine protease p23. The DNA is carried to the nuclear pores in this way, is being released from the capsid remnants and finally translocated into the nucleus as a last step in the transduction process.
One of the limitations of adenoviral vector particles for gene transfer is that many tissues which represent a possible target for therapeutic gene transfer with adenovirus vector particles do not express the CAR-receptor and/or the necessary integrins. Adenovirus vector particles only poorly transduce such tissues even when high doses are applied. Therefore, therapeutic gene transfer is not possible. As examples, the cells of the hematopoietic system, most types of tumor cells, neuronal cells, muscle cells and endothelial cells can be mentioned here. In addition, the use of adenovirus vectors derived from adenovirus isolated from a species other than the one to be treated with the vector might be hampered due to poor transduction efficiencies.
A further limitation of adenovirus vector particles for gene transfer becomes apparent after systemic administration of vector particles into the bloodstream. On the one hand there are interactions with cellular or non-cellular blood components such as erythrocytes, platelets, complement or antibodies which are able to neutralize the vector particles or which are responsible for toxic reactions, and on the other hand there is also CAR-/integrin-mediated uptake into tissues which are not the intended target of the vector particle application. Equally limiting are vector particle interactions with cellular components of the immune system such as Kupffer cells which can lead to neutralization of the vector particles and can induce undesirable side effects.
The current standard of knowledge includes two different strategies whereby one can try to modify the tropism of virus vector particles or one can prevent the undesirable interactions of virus vector particles with, e.g., antibodies.
The first strategy, subsequently referred to as “genetic strategy” consists of a defined genetic modification of solvent-exposed areas of various capsid proteins (e.g., Fiber, protein IX, Hexon). Here one tries to enhance gene transfer into target cells by genetic insertion of peptide ligands into for example the Knob domain of the Fiber protein of Ad5 (Dmitriev, 1998). This genetic strategy shows many concrete disadvantages: (i) there is a significant limitation regarding the size of the ligands to be used, because large ligands interfere with the correct folding of the modified capsid proteins; (ii) it is not possible to predict whether the production of vector particles even after genetic insertion of small ligands is possible because correct protein folding can be disturbed. Furthermore, the structure and biological function of peptide ligands in the context of vector capsid proteins are not predictable and small peptide ligands usually show only limited affinity for the target receptor on the target cell's surface; (iii) substantial genetic modifications, required for successful tropism modification of the adenovirus vector particles, lead to a low yield in vector particle production; (iv) for the production of genetically modified vector particles with new tropism, especially when additional mutations eliminate the interaction with CAR, it is necessary to generate a new production cell line for every ligand—this is a considerable obstacle and inhibits efficient screening for potential ligands; and (v) all genetic procedures are by nature limited to protein/peptide ligands for the tropism modification—peptides containing non-natural amino acids cannot be applied. In addition, other substances than proteins, e.g., steroids, other aromatic compounds or carbohydrates and others cannot be applied.
A second strategy, subsequently referred to as “chemical strategy” consists of unspecific chemical modifications of all solvent-exposed vector particle capsid proteins. Here one uses chemical reactions in order to couple ligands to naturally occuring primary amino groups on the surface of the vector particles. This procedure unspecifically modifies all capsid proteins and is being carried out under oxidative conditions at or slightly above physiological pH (7,4-8,5). (EP0694071B1 and Fisher, 2001) The limitations of the chemical strategy are: (i) the chemical reactions with activated ester groups to form amide bonds or with aldehydes to generate Schiff bases which were used so far, are targeted towards amino groups on the surface of all capsid proteins—single capsid proteins can therefore not specifically be modified; (ii) due to missing specificity for certain capsid proteins, most ligands show a significant cross linking of capsid proteins while coupling—consequently, the natural and for an efficient gene transfer necessary disassembly of the vector particles can severly be impaired after cellular uptake of the modified vector particles; and (iii) the chemical coupling of ligands by activated esters on primary amino groups of the capsid surface under formation of stable amide bonds or by aldehyde groups under formation of Schiff bases is not reversible under biological conditions. Through this, endosmolytic vector particle functions, i.e., functions for the endosomal escape after receptor-mediated endocytosis can be inhibited and can lead to low gene transfer efficiency.
The use of cysteine residues for thiol-specific coupling is a procedure which is being used in various applications. For example Stubenrauch et al. published a paper which describes the coupling of recombinant antibodies to Polyoma-Virus-like Particles (VLP) with the aid of a cysteine residue as an attachment site (Stubenrauch, 2001). A further example for the use of cysteine residues for the formation of covalent bonds is U.S. 2003-219459 A1, which also describes a procedure for the coupling of recombinant proteins to VLPs, whereby the attachment site on the VLP is not a cysteine residue, but the coupling partner is bearing a cysteine residue for attachment to the VLP. Importantly, as opposed to viral vector particles as described here, VLPs are synthetically manufactured particles and are assembled in vitro, i.e., outside living cells and their assembly is totally independent of cellular elements such as transcription factors or chaperone systems und totally independent of cellular processes such as DNA replication. Therefore, they are e.g. not capable of productive infections meaning that they are not able to replicate and multiply under certain conditions.
Recently, Wang et al. described mutants of the Cowpea Mosaic Virus (CPMV), which carry genetically modified solvent-exposed cysteine residues on the capsid surface (Wang, 2002). These mutants were solely produced in order to provide supra-molecular building blocks for chemical synthesis. The production of the plant virus mutant was achieved only under stringent observance of reducing conditions. Deviation from these conditions led to irreversible aggregation and precipitation through formation of interparticular disulfide bridges. The particles of these virus mutants simply serve as an element for chemical synthesis and biological functions were not analyzed.
Upon application of reducing reagents and alkylating reagents on Ad5 virus particles not only the genetically introduced solvent-exposed thiols will be modified, but also the viral cysteine protease p23 of the adenovirus (Greber, 1996). Greber et al. showed for the wildtype adenovirus serotype 2 that after reducing the virus particles by Dithiothreitol (DTT) and subsequent thiolspecific alkylation/esterification by various reagents such as N-ethylmaleimide (NEM) or iodoacetamid (IAA) the viral cysteine protease p23 was converted to its active reduced form and subsequently alkylated. Furthermore, Greber et al. demonstrated that the protease which was alkylated/esterified in this way was inactive and as a consequence the proper disassembly of the capsids after entry of the target cells was hampered in a way that the viral DNA could not be translocated into the nucleus. Greber et al. showed that particle infectivity was reduced 20-fold by alkylating p23 as opposed to non-treated control particles. Furthermore, Greber et al. demonstrated a significant but unspecific alkylation of the cysteine-containing capsid proteins Hexon, Penton base and Fiber as well as of core protein pV with NEM after reduction by DTT. Furthermore, Jornvall and Philipson described a reactive cysteine residue in a solvent-accessible region of the Hexon capsid protein which could be alkylated by maleimide-based reagents (Jornvall, 1980).